Re: Problem in pg_hba.conf - Mailing list pgsql-general

From Mark Dalphin
Subject Re: Problem in pg_hba.conf
Date
Msg-id 385A8AA6.3DB2E616@amgen.com
Whole thread Raw
In response to Problem in pg_hba.conf  ("Mark Alliban" <MarkA@idnltd.com>)
List pgsql-general
Hi,

On  Mon, 13 Dec 1999 10:36:02 -0000, "Mark Alliban" <MarkA@idnltd.com> wrote:

> I have just installed PostgreSQL on SCO OpenServer 5.05 but I am having =
> problems creating a database. My pg_hba.conf file reads:
>
> local        all                                         trust
> host         all         127.0.0.1     255.255.255.255   trust
>
> When I try to create a database, I get:
>
> Postgres$ createdb test
> Connection to database 'template1' failed.
> No pg_hba.conf entry for host localhost, user postgres, database =
> template1
> createdb: database creation failed on test.
>

I look forward to resolution of this problem for I have it too, under special
circumstances.

I run PostgreSQL 6.5.3 on an SGI, running Irix 6.5.  The processor is 64-bit,
but as most the the applications were written for a 32-bit world, SGI supplies a

"compatibility" mode in the compiler which compiles for a 32-bit machine (ie "cc

-n32").  This is what is currently used for creating PostgreSQL under Irix.  For

many reasons, I would like to compile in 64-bit mode ("cc -64"), including that
I have files longer than 4 gigabytes and need to work with the file offsets
(postgreSQL int8 is fine, but Perl is linked to the 32-bit fseek() and it chokes

on the numbers; re-compiling Perl in 64-bit mode is fine, then re-compiling DBI
in 64-bit is fine, but then linking to postgres compiled in 32-bit mode is not
fine.  But I digress...)

When I compile posgresql 6.5.3 in 64-bit mode, I get a clean compile and link.
But when I run it, I get this identical error.  This is using the exact same
setup as I use for the 32-bit compiled postgres. I went so far as to make a
complete removal of the postgres source dir and re-untar, etc to make certain
that the problem was not with some old part lying around. It wasn't. (Also there

was no overlap of old libraries, etc; they were all removed. I used the same
pg_hba.conf for all the tests.).  No amount of playing with pg_hba.conf can make

it work, but removing it caused postgres (64-bit) to complain of its
non-existance.

Any ideas of what might be breaking Postgres when it works with 64-bit numbers
during authentication?

Mark

--
Mark Dalphin                          email: mdalphin@amgen.com
Mail Stop: 29-2-A                     phone: +1-805-447-4951 (work)
One Amgen Center Drive                       +1-805-375-0680 (home)
Thousand Oaks, CA 91320                 fax: +1-805-499-9955 (work)




pgsql-general by date:

Previous
From: Marcin Inkielman
Date:
Subject: [GENERAL] a problem with single char recoding
Next
From: "Culberson, Philip"
Date:
Subject: RE: [GENERAL] How to get timestamped pgsql logging?